Summary:
Ref T2783. With service-oriented calls, we take a larger performacne hit than necessary resolving refs.
Instead of resolving refs over the wire, try to resolve them from the database first. This can resolve almost all refs (commit hashes, branch and tag names).
This can't resolve weird refs like `master~50`, and obviously can't resolve invalid refs. In those cases we'll go back to the old logic, call `diffusion.resolverefs`, and end up with the right result.
Test Plan:
- Browsed repositories in Diffusion.
- Verified that service repositories no longer make unnecessary `diffusion.resolverefs` calls for common refs (branch names, commit hashes).
- Resolved refs like `master~50`, saw call to underlying VCS and correct result.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T2783
Differential Revision: https://secure.phabricator.com/D11476
Summary: Fixes T6395. Ref T6350. I guess I missed this code spot in prior testing / I definitely didn't run an empty commit through it. Works now though.
Test Plan: made an empty commit and observed stuck importing status and errors in phd log. applied patch and commit successfully imported with no errors. made another empty commit and it imported as well
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin, epriestley
Maniphest Tasks: T6350, T6395
Differential Revision: https://secure.phabricator.com/D10746
Summary:
I am not sure how valuable this is *as is* - I think it needs different explanations for what happened in mercurial or subversion? I do not know what those explanations are.
Made an error in D10485 - the $hashes that were saved is an array of objects, so it ends up turning into garbage via the wonders of serialization and de-serialization. Fix that by explicitly saving the tree hash.
I would like to make this work for the other VCS types we support, add the "undo / nope" button and call it fixed.
Ref T3686.
Test Plan: clicked "explan why" and saw why
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: epriestley, Korvin
Maniphest Tasks: T5693, T3686
Differential Revision: https://secure.phabricator.com/D10489
Summary: Fixes T4387.
Test Plan: Setup a mercurial repository for rabbitmq-server. Browsed around it and things looked good.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: epriestley, Korvin
Maniphest Tasks: T4387
Differential Revision: https://secure.phabricator.com/D10380
Summary:
Fixes T5839. If a repository has been force pushed and garbage collected, we might have a ref cursor in the database which still points at the old commit (which no longer exists).
We'll then run a command like `git log <new hash> --not <old hash>` to figure out which commits are newly pushed, and this will bomb out because `<old hash>` is invalid.
Instead, validate all the `<old hash>` values before we try to make use of them.
Test Plan:
- Forced a repository into a bad state by mucking with the datbase, generating a reproducible failure similar to the one in T5839.
- Applied patch.
- `bin/repository update <callsign> --trace` filtered the bad commit and put the repository into the right state.
- Saw new commits recognized correctly.
- Ran `bin/repository update <callsign>` for a Mercurial and SVN repo as a sanity check.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T5839
Differential Revision: https://secure.phabricator.com/D10226
Summary:
Fixes T5304. Mercurial features a "{branches}" template keyword, documented as:
```
branches List of strings. The name of the branch on which the
changeset was committed. Will be empty if the branch name
was default.
```
At some time long in the past, I misinterpreted this to mean "list of branches where the branch head is a descendant of the commit". It is more like "list of zero or one elements, possibly containing the name of the branch the commit was originally made to, if that branch was not 'default'".
In fact, it seems like this is because a //very// long time in the past, Mercurial worked roughly like I expected:
> Ages ago (2005), we had a very different and ultimately unworkable
> approach to named branches that worked vaguely like .hgtags and allowed
> multiple branch names per revision.
http://marc.info/?l=mercurial-devel&m=129883069414855
This appears to be deprecated in modern Mercurial (it's not in the modern web documentation) although I can't find a commit about it so maybe that's just a documentation issue.
In any case, `{branches}` seems to never be useful: `{branch}` provides the same information without the awkward "default-if-empty" case.
Switch from `{branches}` to either `{branch}` (where that's good enough, notably in the hook engine) or `(descendants(%s) and head())`, which is equivalent to `--contains` in Git.
This fixes pushing to branches with spaces in their names, and makes the "Branches" / "Contains" queries moderately more consistent.
Test Plan:
- Pushed to a Mercurial branch with a space in it.
- Viewed list of branches in a Mercurial repository.
- Viewed containing branches of a Mercurial commit in Diffusion.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T5304
Differential Revision: https://secure.phabricator.com/D9453
Summary: Ran `arc lint --apply-patches --everything` over rP, mainly to change double quotes to single quotes where appropriate. These changes also validate that the `ArcanistXHPASTLinter::LINT_DOUBLE_QUOTE` rule is working as expected.
Test Plan: Eyeballed it.
Reviewers: #blessed_reviewers, epriestley
Reviewed By: #blessed_reviewers, epriestley
Subscribers: epriestley, Korvin, hach-que
Differential Revision: https://secure.phabricator.com/D9431
Summary: Ref T4327. At some point these two very similar classes got introduced. Collapse `DiffusionBranchInformation` into the nearly identical `DiffusionRepositoryRef`, which enjoys slightly more generality and support.
Test Plan: Viewed branch overview and detail pages. Ran `repository refs` and `repository discover`. Grepped for removed symbols.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T4327
Differential Revision: https://secure.phabricator.com/D8002
Summary: Ref T4195. Ref T2783. We have an old-school implementation of this; move it into a LowLevel query and make callers all run through Conduit. I need the LowLevel query for hooks, to implement an "is merge commit" Herald rule.
Test Plan:
- Ran query via Conduit for SVN, Mercurial, Git.
- Parsed a commit which closed a revision, attach/closed worked correctly.
- Browsed Diffusion.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T4195, T2783
Differential Revision: https://secure.phabricator.com/D7808
Summary: Ref T4195. I need to query commit metadata to figure out which revision a commit is associated with. Move this out of the MessageParser so the code can be called from the HookEngine.
Test Plan: Used `reparse.php` to reparse a variety of SVN, Mercurial and Git commits. Used `var_dump()` to verify sensible fields were returned.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T4195
Differential Revision: https://secure.phabricator.com/D7805
Summary: Ref T4195. I need this for the Herald pre-commit rules, and it generally simplifies things.
Test Plan: Used `reparse.php` plus `var_dump()` to inspect refs in Git, Mercurial and SVN repos. They all looked correct and reparsed correctly.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T4195
Differential Revision: https://secure.phabricator.com/D7804
Summary: Ref T4195. Even though we use `svnlook` in the hook itself, I need this query elsewhere, so provide it and merge the classes into one which does the right thing.
Test Plan:
- Used `reparse.php` to reparse messages for Git, SVN and Mercurial commits, using `var_dump()` to examine the commit refs for sanity.
- Used `reparse.php` to reparse changes for an SVN commit.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T4195
Differential Revision: https://secure.phabricator.com/D7800
Summary: Ref T4195. Same as D7793, but for mercurial. (As usual, SVN needs some goofy nonsense instead, so the next diff will just make this field work.)
Test Plan: Ran `reparse.php` on Git and Mercurial commits, var_dump'd the output and it looked correct.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T4195
Differential Revision: https://secure.phabricator.com/D7795
Summary: Ref T4195. I need to issue this command from the pre-commit hook to get commit bodies for hooks.
Test Plan: Ran `reparse.php --message --trace` and dumped the $ref, which looked correct.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T4195
Differential Revision: https://secure.phabricator.com/D7793
Summary:
Ref T615. Ref T4237. With `--debug`, Mercurial will echo an "ignoring untrusted configuration option" warning **to stdout** if `.hgrc` has the wrong owner.
However, we need `--debug` to make `{parents}` usable, at least until the patches I got into the upstream are widely deployed. So after getting `--debug` output, strip off any leading warnings.
These warnings should always be in English, at least, since we set `LANG` explicitly.
Test Plan: Unit tests. @asherkin, maybe you can confirm this? I can't actually get the warning, but I think my `hg` in PATH is just a bit out of date.
Reviewers: asherkin, btrahan
Reviewed By: asherkin
CC: asherkin, aran
Maniphest Tasks: T615, T4237
Differential Revision: https://secure.phabricator.com/D7784
Summary:
Ref T1493. Diffusion has some garbagey behavior for things we can't resolve. Common cases are:
- Looking at a branch that doesn't exist.
- Looking at a repository with no branches.
- Looking at a commit that doesn't exist.
- Looking at an empty repository.
In these cases, we generally fatal unhelpfully. I want to untangle this mess.
This doesn't help much, but does clean things up a bit. We currently have two separate query paths, "stablecommitname" and "expandshortcommit". These are pretty much doing the same thing -- taking some ref like "master" or "default" or a tag name or part of a commit name, and turning it into a full commit name. Merge them into a single "resolverefs" method.
This simplifies the code a fair bit, and gives us better error messages. They still aren't great, but they're like this now:
Ref "7498aec194ecf2d333e0e2baddd9d5cdf922d7f1" is ambiguous or does not exist.
...instead of just:
ERR-INVALID-COMMIT
Test Plan: Looked at Git, Mercurial and Subversion repositories that were empty and non-empty. Looked at branches/heads. Tried to look at invalid commits. Looked at tags. All of this still works, and some behaviors are a bit better than they used to be.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T1493
Differential Revision: https://secure.phabricator.com/D7484
Summary: Ref T1493. Consolidate these a bit; they might need some more magic once we do `--noupdate` checkouts. Mostly just trying to clean up and centralize this code a bit.
Test Plan: Viewed and `bin/repository discover`'d Mercurial repos with and without any branches.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T1493
Differential Revision: https://secure.phabricator.com/D7480
Summary: Ref T2230. This cleans up D7442, by using `git for-each-ref` everywhere we can, in a basically reasonable way.
Test Plan:
In bare and non-bare repositories:
- Ran discovery with `bin/repository discover`;
- listed branches on `/diffusion/X/`;
- listed tags on `/diffusion/X/`;
- listed tags, branches and refs on `/diffusion/rXnnnn`.
Reviewers: btrahan, avivey
Reviewed By: avivey
CC: aran
Maniphest Tasks: T2230
Differential Revision: https://secure.phabricator.com/D7447